Cohorts are a way to identify actors who did a thing.
Interana allows you to define cohorts of actors based on any attribute of those actors, including their journey as reflected in the events stored for them in Interana.
Use cohorts in filters to drill down and ask targeted questions about subsets of your actors.
These cohorts, like everything else in Interana, are evaluated entirely at read time, so you can get immediate results based on new segments at any time without any reprocessing of historical data. This enables both demographic and behavioral segmentation entirely on-the-fly.
Time can be used in multiple ways in cohorts: specific time windows, relative time (rolling windows), long time periods to cover entire data range.
Best practices for cohorts
- Inclusive logic: cohort membership is determined at the event level, so making the logic inclusive (e.g., actors who DID do a particular action) avoids unexpected actors as cohort members.
- Exclusive logic: you can use “not in cohort” as a way of excluding cohort members from a query.
- Time ranges: pay special attention to the Time associated with the cohort. We recommend including this in the name of the cohort.